It is difficult to organize the entire content to be voice-inputted in the brain in advance and make the utterance perfect. Even in the consciousness of voice input, the utterance tends to be made with a voice that is highly likely to fail in voice recognition during the voice input. For example, when an utterance is made using hesitations, slips of the tongue, unnecessary words inserted unconsciously, words that the voice recognition is weak in and that tend to fail in recognition, etc., such as “uh, er” and “uh-huh” it is highly likely that the voice recognition will fail.
Thus, it is difficult to acquire an intended text from an inputted voice, and hence voice recognition technology to enhance the accuracy of the text acquired from a voice is under continues investigation (e.g. Patent Literature 1 etc.).